[[alternative]]The Application of Fuzzy Neural Network to the Greenhouse Climate Control

碩士電機工程學系[[abstract]]本論文採用目前常用的模糊類神經網路(Fuzzy Neural Network；簡稱FNN)，來控制一個複雜的溫室環境，並利用可程式控制器（Programmable Logic Controller；簡稱PLC）配合個人電腦上的Labview虛擬儀錶軟體程式，進行監視與控制，最後透過網際網路的連線，讓整個監控系統的效能更趨於完善。透過模糊類神經網路的控制，我們不需要去分析每一種會影響溫室環境的因素，讓我們可以更容易的去控制一個系統；而利用Labview虛擬儀錶所設計出來的人機界面，能讓操作員更容易的了解目前系統的狀況，以便進行控制。 在控制方面，由於類神經網路不需要分析複雜的系統，只需要提供訓練樣本，建立訓練資料庫後，並讓模糊類神經網路去學習，待學習完成後，它便具有控制系統的能力；而且因為模糊類神經網路擁有聯想、預測的能力，就算目前的輸入值沒有在訓練的資料庫內，它也可以做出適當的反應，讓整個控制的性能更趨向完善，讓整個溫室被控制在一個適合農作物生長的環境，這也是本論文最終的應用目的。[[abstract]]This thesis adopts fuzzy neural network ( FNN ) to control a complicated greenhouse climate. It is designed by programmable logic controller to go with human-machine-interface ( HMI ) in PC for monitoring and controlling. The whole supervisory control system will finally make the best efficiency by connecting Internet. It makes us control the system easier and we don’t need to analyze each factor of affecting greenhouse climate by using FNN control. To use the HMI that was designed by Labview will make the operators understand the system status and control the situation easier. In the aspect of controlling, FNN doesn’t need to analyze the complicated system and the engineer just need to offer FNN training samples for studying. After studying, it has the ability of controlling system. It still can make proper responses even though the input samples are never trained by FNN because it has the ability of association and forecast. Then it will make the whole greenhouse that is controlled under the suitable climate for those growing crops